The recent release of the 2.5 Linux kernel sparked some heated discussion in the open source community, specifically surrounding the ability of Linux's creator, Linus Torvalds, to keep up with the growing number of patches and fixes to the kernel.
Some of the complaints in the community are that new versions of the kernel are released before older versions have completely stabilized. With these rapid releases, some fear the Linux market will become too fragmented and the developers will become frustrated and dissatisfied. Some even feel these public arguments are harmful to the open-source movement.
One of the problems with open-source development is that it's so... well, open. On the commercial side, most of these same topics -- resource issues, inclusion of particular patches and features into a new product -- are kicked around in offices around the world. But these discussions occur in private, away from the public eye. When it comes to commercial products, we generally don't get to the failures of early integration efforts and are not privy (and thus can't contribute) to decisions on features. We simply never hear about any of this.
Open-source development is, by its very nature, a public endeavor. As such, we tend to hear about every little disagreement that occurs.
Fortunately, enterprises can generally sidestep these issues by sticking with major, proven Linux releases. For instance, distribution vendors such as RedHat, SuSE and Caldera apply their own patches to the kernel before release to the public. It's in their best interests to ensure a quality product. Vendor-supplied distributions are arguably as stable, and probably more so, than anything Microsoft has released in the past few years.
People who are concerned about the "fragmentation" of Linux should realize that fragmentation is already here. But this fragmentation is not necessarily the big deal that alarmists have made it out to be. Commercial Linux vendors lag behind and wait for a kernel that, if not already stable, can be stabilized and then provided to the community at large. Those who want bleeding-edge-omigod-this-hurts-to-install experiences are welcome to experimental kernels and development builds. But don't equate the availability of a new kernel as a sign from on high that you must upgrade now. That's a Microsoft mentality, and it simply doesn't make sense. In fact, if Linux weren't open source you wouldn't even hear about these releases.
Should Linus slow down? Should he get some more help? That's up to him. It's his baby, his creation. As a hard-core developer, I grok Linus and his reluctance to offer up to the hoards his creation. It isn't his dream that Linux grow into a corporate-class operating system -- that's someone else's desire, and someone else's effort. Is this bad for open source? Not at all. Open source is what gives the RedHats and Calderas of the world the ability to go ahead and patch, stabilize and modify the kernel and provide a corporate-class incarnation of Linux. And that's a good thing.
If you're using -- or thinking of using -- Linux in the enterprise, just stay away from experimental kernel releases for production environments. They really aren't intended for public consumption; they simply allow people to test, try out and play with new features and enhancements in the kernel. Stick with your distribution, and if you're interested in new features, wait for an integrated and upgraded distribution with the newest kernel.
RSS
